SSINS No.: 6835 
                                                         IN 87-17       

                                UNITED STATES
                        NUCLEAR REGULATORY COMMISSION
                    OFFICE OF INSPECTION AND ENFORCEMENT
                           WASHINGTON, D.C. 20555

                                April 7, 1987

Information Notice No. 87-17:   RESPONSE TIME OF SCRAM INSTRUMENT VOLUME 
                                   LEVEL DETECTORS 

Addressees: 

All General Electric boiling water reactor (BWR) facilities holding an 
operating license or a construction permit. 

Purpose: 

This notice is to alert addressees of the potential for long response times 
in level detection instruments used to measure the water volume in the scram 
discharge instrumented volume of BWRs. These instruments are intended to 
cause a timely scram prior to filling of the scram discharge volume (SDV). 
Recipients are expected to review the information for applicability to their 
facilities and consider actions, if appropriate, to preclude similar 
problems occurring at their facilities. However, suggestions contained in 
this information notice do not constitute NRC requirements; therefore, no 
specific action or written response is required. 

Background 

On June 28, 1980 during a routine shutdown of the Browns Ferry Unit 3 
reactor, a manual scram from about 36 percent power failed to insert 
approximately 40 percent of the control rods. The root cause was isolated to 
a problem with the SDV header. The followup to this event at other BWRs 
revealed a number of deficiencies that were discovered with the SDV headers.

The corrective measures to this problem are divided into a short-term 
program and a long-term program. The short-term actions were implemented by 
IE Bulletins 80-14 and 80-17 and their supplements. The long-term program 
was addressed by a Generic Safety Evaluation Report (SER) dated December 1, 
1980. The objective of the long-term program was the improvement of the SDV 
design. (SDV design details vary from plant to plant.) One of areas of 
concern addressed by the long-term program was that the reliability of the 
float switches in the instrumented volumes needed to be improved. The 
recommended solution was the addition of redundant and diverse water level 
sensors. 

Description of Circumstances: 

More recently, the NRC reviewed the licensee's corrective actions taken at 
the Browns Ferry units as a result of the earlier event and identified a 
potentially generic concern with the response time of the level detectors 
for

8704030045
.

                                                            IN 87-17     
                                                            April 7, 1987 
                                                            Page 2 of 2  

the instrumented volume. As an interim solution to issues raised by the 1980 
event, the licensee installed a circuit to scram the reactor upon sensing 
low air header pressure. As a long-term fix, the licensee then installed 
differential pressure (DP) detectors as a diverse means of sensing a full 
instrumented volume and causing a scram. The DP detectors back up the float 
switches in the instrumented volume that also can cause a scram. 

Testing by the licensee showed a response time of about 70 seconds for the 
installed DP detectors in the instrumented volume. This was unacceptable 
(see discussion). The licensee replaced the DP detectors with 
resistance-temperature detectors (RTDs). Comparison of the response time of 
the RTDs with the float switches showed that the float switches had a 
response time of about 20 seconds. This was also unacceptable. The licensee 
has consequently upgraded the scram on low air header pressure 
instrumentation to a permanent installation that is environmentally 
qualified and covered by technical specifications. 

Discussion: 

Although the SER did address the minimum capacity of the SDV, the 
potentially generic concern is that the SER did not address the response 
time of the level detection instruments in the instrumented volume. There is 
a direct correlation between the response time of these instruments and the 
capacity of the SDV. 

Under degraded air conditions, the scram outlet valve for each control rod 
drive may leak as much as 5 gpm to the SDV without rapid rod motion. The 
extent and timing of water accumulation in the SDV can slow or stop rod 
insertion on scram. (GE has experimentally shown that if water has 
accumulated to about 85 percent or greater of the SDV header volume, scram 
is effectively stopped.) The instrumented volume level instrumentation are 
intended to generate a timely scram signal such that the reactor is scrammed 
before water leaking by the scram outlet valves fills the SDV. 

No specific action or written response is required by this information 
notice. If you have questions about this matter, please contact the Regional 
Administrator of the appropriate NRC regional office or this office. 


                         Edward L. Jordan Director
                         Division of Emergency Preparedness
                           and Engineering Response
                         Office of Inspection and Enforcement 

Technical Contact:  Eric Weiss, IE
                    (301) 492-9005 

                    Charles Patterson, NRC Resident Inspector
                    Browns Ferry Nuclear Power Plant
                    (205) 729-6196

Attachments:
List of Recently Issued IE Information Notices
.